Single-Frequency High-Power Continuous-Wave Oscillation at
1003 nm of an Optically Pumped Semiconductor Laser
M. Jacquemeta, M. Domenecha, J. Dionb, M. Strassnerb,
G. Lucas-Leclina, P. Georgesa, I. Sagnesb and A. Garnachec
Laboratoire Charles Fabry de l’Institut d’Optique
UMR 8501, Centre Universitaire, Bat. 503, 91403 Orsay cedex (France)
gaelle.lucas-leclin@iota.u-psud.fr
a
b
Laboratoire de Photonique et de Nanostructures
CNRS UPR20, Route de Nozay, 91460 Marcoussis (France).
c
Centre d’Electronique et de Micro optoélectronique de Montpellier, CNRS UMR 5507,
Universite Montpellier II, Place E. Bataillon, 34095 Montpellier cedex 5 (France).
ABSTRACT
This work reports single-frequency laser oscillation at  = 1003.4 nm of an optically pumped external cavity
semiconductor laser. By using a gain structure bonded onto a high conductivity substrate, we demonstrate both
theoretically and experimentally the strong reduction of the thermal resistance of the active semiconductor medium,
resulting in a high power laser emission. The spectro-temporal dynamics of the laser is also explained. Furthermore, an
intracavity frequency-doubling crystal was used to obtain a stable single-mode generation of blue ( = 501.5 nm) with an
output power around 60 mW.
Keywords: Vertical External Cavity Semiconductor Emitting Laser (VECSEL), Thermal Management, Single
Frequency, Non-linear conversion.
1. INTRODUCTION
Optically pumped Vertical-External-Cavity Surface Emitting Lasers (VECSEL’s) combine the approaches of diodepumped solid-state lasers and engineered semiconductor lasers, generating both circular diffraction limited output beams
and high average powers [1,2]. However the poor thermal conductivity of III-V materials might prevent an efficient heat
removal from the active region, which results in an excessive heating of the semiconductor structure incompatible with
high power emission. Mainly two solutions have been described in the literature: first growing the structure upside down
and removing the GaAs substrate to dissipate the heat directly through the Bragg mirror [1,3]; or secondly removing the
heat through the top of the semiconductor structure by bonding it to a material of high thermal conductivity and good
optical quality [2,4,5]. In this work we propose solid-liquid inter-diffusion (SLID) bonding to transfer
temperature sensitive components without introducing additional strains. This technique benefits from a relatively low
bonding temperature (~200° C) and a good temperature stability of the bond since the melting point of the AuIn2 alloy is
480° C. Further high temperature processing steps, such as plasma enhanced vapor phase deposition of dielectric
materials, are thus possible without dissolving the AuIn2 bond.
Output powers up to 500 mW in a single transverse and longitudinal mode have been obtained at 1003 nm from an
optically pumped ½ VCSEL bonded onto a SiC heatspreader. Furthermore, intracavity second harmonic generation is
reported at the blue-green wavelength of 501.5 nm, which corresponds to a hyperfine transition of the molecular iodine
(127I2) of great interest in metrological applications as well as in high-resolution spectroscopy [6,7].
1/11
Proceedings Photonics Europe Strasbourg 2006
2. ½ - VCSEL STRUCTURE FOR THERMAL MANAGEMENT
2.1 Gain structure design and fabrication
The L=1010 nm ½ -VCSEL designed for p=808 nm pumping used in the present work was grown by MOCVD on a
350 µm – thick GaAs (100) substrate [8]. The multilayer Bragg mirror is made with 27.5 pairs of GaAs/AlAs. The
3.75 -thick active region consists of NQW = 5 compressively strained LQW=7 nm InGaAs/GaAs quantum wells (QWs)
distributed among the 7 optical standing-wave antinode positions with a distribution function 1-1-1-0-1-0-1 (starting
from the top surface) such as the carrier density remains almost equal in all QWs. We have optimized the number of
QWs to obtain a low laser threshold and a large differential gain at 300 K with 1% of total optical cavity losses [9]. The
low number of QWs, as compared to other published designs, avoids the need to grow strain-compensating layers
between the QWs [1-4]. On top of the gain region a 30 nm AlAs confinement layer and an 8 nm caping layer were
added.
A second structure, with exactly the same layers but an AlGaAs etch-stop layer between the GaAs substrate and
the Bragg mirror, was grown on reverse order on a GaAs substrate for further processing. This ½-VCSEL was bonded
onto a 280-µm thick Si substrate (thermal conductivity of s1 =150 W/m/K) on the Bragg mirror side by solid-liquid
inter-diffusion (SLID) bonding with AuIn2 (figure 1) [10]. A 30 nm-thick Ti layer was first evaporated on both the ½ VCSEL and the SiC heatspreader to improve adhesion and to prevent diffusion. Next, a 150 nm –thick Au layer was
deposited on the SiC substrate. On top of the active structure we added Au (150 nm) and In (600 nm) layers; then a
20 nm-thick Au layer was deposited to prevent the oxidation of the Indium layer prior to bonding. The bonding
temperature was 200° C, which was a trade-off between the diffusion speed and the strain introduced by the bonding
process due to the different thermal expansion coefficients of both materials. A pressure of about 250 kg / cm² was
applied to provide a good surface contact of the two wafers while the temperature was ramped to 200° C and maintained
at this temperature for two hours. The pressure was kept constant during the cooling-down process. The final thickness of
the AuIn2 bonding layer was found to be 0.9 µm. Then both mechanical polishing and wet selective etching at room
temperature was utilized to remove the GaAs substrate. Finally, a Si3N4 anti-reflection coating was deposited by plasma
enhanced chemical vapor phase deposition (PECVD) at high temperature (280° C) to protect the top surface from
oxidation and to weaken the residual etalon resonances of the semiconductor structure.
Figure 1: Description of the SLID bonding process of the ½ VCSEL structure onto a SiC heatspreader.
The reflectivity and photoluminescence (PL) spectra of the bonded structure are shown in Figure 2,
demonstrating an optimized design with a good matching between the central wavelength of the Bragg mirror and the
QW gain. Neither a PL broadening nor a PL wavelength shift have been observed for the bonded structure,
demonstrating that our AuIn2 bonding technique does not modify the optical properties of the QWs. With a residual air-
2/11
Proceedings Photonics Europe Strasbourg 2006
structure interface reflection of about 6%, and a calculated pump absorption at P = 808 nm of nearly 70 % in the barrier
layers, the fraction of incident pump power absorbed in the active region is evaluated to Ap=66%. The remaining 30%
pump power is absorbed in the GaAs layers of the Bragg mirror.
.
Figure 2: Reflectivity of the half-VCSEL bonded onto the SiC substrate (left; black curve : with AR
coating, grey curve : without AR coating) and photoluminescence spectrum (red curve, right) under
low-power excitation before the AR coating deposition. The residual etalon effect of the ½ -VCSEL
remains visible on the reflectivity spectrum around 1.0 µm.
2.2 Theoretical and experimental evaluation of the thermal properties of the gain structures
The heat diffusion in the whole semiconductor structure has been theoretically studied assuming a perfect metal bonding
between the Bragg mirror and the SiC substrate. The pump absorption was supposed to decrease exponentially through
the active region and the Bragg mirror; the pump beam transverse profile was top-hat. We have evaluated the average
thermal resistance Rth of the ½ -VCSELs over the laser beam area, measured between the top surface of the Peltier cooler
and the QWs. The values of Rth are given relatively to the incident pump power. Assuming an internal quantum
efficiency i of 80% for the QWs and of 10% for the GaAs in the Bragg mirror, 50% of the incident pump power is
converted to heat.
On the one hand, we used an analytical three-layer model (active region/susbtrate/copper block) in which the
substrate itself is perfectly bonded onto an infinite copper block (thermal conductivity of 4 W/m/K) [11]. The active
region consists of the QWs and the Bragg mirror layers, with a total thickness e v=5.45µm. A mean value of
v=32 W/m/K was considered as its thermal conductivity. We have evaluated that for pump beam waist radii larger than
the critical value rc= ev (s/v), the major contribution to the temperature increase in the QW layers arises from the
substrate. Then the heat flow in the ½–VCSEL is purely 1D, and the temperature rise only depends on the pump power
density. These critical pump radii are rc= 7.7 µm and 83.4 µm for the GaAs and SiC substrates respectively. On the other
hand, the thermal resistance has been evaluated with a finite-element numerical simulation (FEMLAB). A 10-mm lateral
width of the sample was considered. Furthermore, a 50 µm-thick heat-paste layer (thermal conductivity of 2W/K/m)
embedded between the substrate and copper block and the substrate could be taken into account The thermal resistances
evaluated from both the numerical and the analytical models are compared in Figure 3. The analytical model, which
mainly neglects the heat-paste layer contribution, slightly underestimates the thermal resistance from 2 to 60% in the
range of pump radius investigated here (30-450µm). It is thus a simple and efficient tool for the evaluation of the thermal
resistance of the structures.
3/11
Proceedings Photonics Europe Strasbourg 2006
Laser Thermal Impedance (K/Winc)
1000
Analytical 3-layer model
Numerical model, with 50µm heatpaste
Numerical model, without 50µm heatpaste
Experimental evaluation
100
10
1
0,1
0
100
200
300
400
500
Pump Radius (µm)
Figure 3: Theoretical evaluations (lines: analytical, circles: numerical) of the thermal resistance of the
½ -VCSEL structures on respectively SiC and GaAs substrates, vs the pump radius, and comparison
with the experimental values (crosses).
These simulations have been compared to the experimental evaluation of the thermal resistance of the two samples under
laser operation close to threshold, through the measurement of the emission wavelength shift with the incident pump
power and the heatsink temperature. For a pump waist radius of 100 µm, a thermal resistance of 6 K/W inc was measured
for the structure bonded onto SiC, which is in relatively good accordance with the theoretical value of 5.6 K/Winc. This
value is much lower than the thermal resistance measured under the same experimental conditions for the structure on the
GaAs substrate, which was about 50 K/Winc. Finally, the temperature at the top surface of the ½-VCSELs has been
measured under cw optical pumping in the spontaneous emission regime (no laser cavity) with an infrared camera
operating in the 8-12µm range [12]. The lateral resolution of the mapping was 60µm, thus only suited for large pump
beam waists. With a maximum temperature rise limited to 15°C for a 4 W pump beam focused on a 250 µm radius waist,
the structure bonded onto SiC obviously demonstrates the lower thermal resistance (figure 4). Furthermore these maps
show the good homogeneity of our bonding process.
Figure 4: Temperature maps of the ½ -VCSEL structures under optical pumping (= 808 nm, Pp = 4
W, Wp = 250 µm) without laser operation.
4/11
Proceedings Photonics Europe Strasbourg 2006
III. INFRARED LASER CHARACTERIZATION
3.1 Experimental set-up
The VECSEL system was set up in a simple and compact cavity configuration as it is shown in Figure 5. The halfVCSEL was mounted onto a copper heatsink to control the temperature of the device with a Peltier module. The laser
was pumped with a continuous wave (cw) fiber-coupled p = 808 nm laser diode delivering up to 7 W focused on
approximately a 100 µm radius spot with an incidence angle of 20°. The external cavity was simply formed by the
VCSEL and a concave dielectric mirror with a reflectivity of R oc= 98.8 % at L=1010 nm and a curvature radius of 100
mm. The overall cavity length was around Lc=95 mm, leading to a cavity Free-Spectral-Range (FSR) of ~ 5 pm.
Figure 5: Experimental set-up for high-power laser emission in the infrared.
3.2 High power laser characterization: experiment and theory
In the compact two-mirror cavity and without any intracavity element, an output power of 1.7 W at the maximum
incident pump power of 7 W was obtained with the ½ - VCSEL bonded on SiC at 283 K. On the contrary, no laser
operation was achievable with the structure on GaAs in the same cavity and pump conditions, and the maximum output
power was only 100 mW for a pump waist radius of 50 µm. This demonstrates once more the effectiveness of the
bonding technique on a high conductivity substrate for high power laser emission. Further laser investigations will thus
only be reported with the ½ - VCSEL bonded onto SiC: the incident pump power at threshold was 0.8 W in cw operation
at 283 K, corresponding to an incident pump density of I incth=2.2kW/cm2, and the CW slope efficiency was about 32 %.
The laser oscillation occurs around 1009 nm at the maximum pump power, with a cw spectral bandwidth of nearly 2 nm.
We believe that this linewidth is limited by the QW's gain jitter induced by the pump power fluctuations.
These experimental results have been compared to a theoretical evaluation of the VECSEL behavior based on
the following equations:
I inc  J current 
gth  g0 ln
 e  i
N QW hc
 p Ap e
th
J current
T  (1  Roc )
 i
J tr
2c NQW LQW
p
1  Roc
Ap 
L
(1  Roc )  Ti
(Eq.1)
(Eq.2)
(Eq.3)
Equations (Eq. 1) and (Eq. 2) respectively sets the relation linking the incident pump power density I inc to the current
density per QW Jcurrent (in A/cm2), and the material gain at threshold gth to the cavity parameters. The values for g0
5/11
Proceedings Photonics Europe Strasbourg 2006
(1200 cm-1) and the Jtr (50 A/cm2), which make the connection between the maximal material gain and the current
density per QW, are deduced from reference 9 for InGaAs/GaAs quantum wells. The longitudinal confinement factor c
is 2 and the intracavity losses Ti are evaluated to 0.2%. The external quantum efficiency e at the laser output, assuming
a pump-to-laser beam recovering integral close to unity, is given by (Eq. 3). Finally, the evolution of the threshold with
temperature is theoretically taken into account through the characteristic temperature T 0, evaluated to 60 K in these
experiments. Figure 6 shows the very good agreement between the theoretical evaluation of the output power with the
incident pump power, and our experimental results. An actual output power limited to 1.8 W is predicted for an incident
pump power of 7.9 W and a 100 µm radius pump waist.
These simulations have been used to estimate the maximum output power which could be extracted from our ½VCSEL at T = 283 K, with an optimization of both the incident pump power and the pump beam waist. In these
calculations the thermal resistance of the structure was evaluated with a 50 µm-thick heatpaste layer between the copper
heatsink and the substrate (see figure 3). Obviously the low thermal conductivity of GaAs limits the available laser
power to 0.1 W (figure 6, inset). On the contrary an output power as high as 6.7 W should be obtained with a 30 W
incident pump beam focused onto a 260 µm radius waist with the structure bonded on SiC.
1.5
10
Output laser power (W)
Output Laser Power (W)
2
1
0.5
SiC
Pp=30 W
1
0.1
GaAs
Pp=0.5 W
0.01
0
100
200
300
400
Pump waist radius (µm)
500
600
0
0
2
4
6
Incident Pump Power (W)
8
10
Figure 6: Experimental (full circles) and theoretical (line) output laser emission from the structures
bonded onto SiC at T = 10°C, for a pump beam waist radius Wp = 100 µm. Inset : theoretical
optimization of the output laser power vs the pump waist radius for the ½ -VCSEL on GaAs and the
one bonded onto SiC ; the incident pump power is set for maximum laser emission of each structure; T
= 10°C, output coupler T = 1.2%.
10
3.3. Single-frequency laser operation
The insertion of a solid 50 µm-thick low-finesse etalon in the compact cavity set-up described in figure 5 results in a
narrow-line laser emission tunable from ~ 998 nm to ~ 1010 nm by rotating the etalon and changing the pump position
on the ½ -VCSEL structure. The maximum output power was 1.3 W at 1006.5 nm for an incident pump power of 7 W;
the reduction of the output power by about 30 % as compared to the previous results is due to the additional losses
introduced by the solid etalon. On the entire incident pump power range the emission wavelength remains strongly
6/11
Proceedings Photonics Europe Strasbourg 2006
controlled by the etalon with a spectral shift less than 0.2 nm. The beam quality factor M2 relating the spatial beam
properties to an ideal Gaussian beam was about 1.3 at the highest pump power. The linear polarization of the laser beam
was aligned along the [110] semiconductor crystal axis, due to a slight gain dichroism between [110] and [1-10] crystal
axis.
A true single-frequency operation was obtained at 1003 nm for an output power limited to 0.5 W at the
maximum incident pump power. That further decrease of the output power can be mainly attributed to the alignments of
the cavity and the etalon to force a stable single-frequency laser line. The single-frequency laser spectrum was analyzed
with a 50 mm confocal scanning Fabry-Perot interferometer with a FSR of around 1.5 GHz (see inset in figure 7). The
side-mode suppression ratio is at least 20 dB, but the scanning Fabry-Perot apparatus function limits the measurement of
the laser linewidth. However in this configuration, the laser linewidth should be limited by acoustic and/or thermal noise
at the kHz level [4,13]. No stabilization of the laser emission frequency has been carried out.
.
Figure 7: CW laser output power vs the incident pump power with the intracavity 50 µm-thick etalon
for the structure bonded onto SiC; open squares: narrow spectrum laser emission at 1003 nm; full
squares: true single-frequency operation at 1003 nm; inset: scanning Fabry-Perot trace for the emission
at 1003.4 nm demonstrating single-frequency operation with the FP etalon inside the laser cavity. The
½ -VCSEL temperature is 10°C.
It has recently been demonstrated both theoretically and experimentally that a single-frequency operation is
achievable in short cavity (Lc < 100 mm) VECSEL without any intracavity filter, thanks to their ideal multimode
homogeneous gain and the fact that non-linear mode coupling is negligible [13,14]. Indeed, though at the very beginning
of the laser operation several thousands of longitudinal cavity modes are amplified, the laser collapses to a single mode
after a characteristic time tc. But the emission would remain multimode if any process (gain or mirror jitter, thermal
fluctuations…) disturbs the laser dynamics on a time scale shorter than t c. The characteristic time tc necessary to achieve
a single-mode laser operation after a strong perturbation arises (100% gain modulation e.g.) is equal to the time t g within
which the non-stationary laser bandwidth L, given by (Eq. 4), is equal to one FSR of the laser cavity:
 L  g, f

7/11
ln2
 0t g
(Eq.4)
Proceedings Photonics Europe Strasbourg 2006
Characteristic time tc (ms)
In this equation, tg is the generation time elapsed since the beginning of the laser emission, 0 is the broadband cavity loss
rate (proportional to 1/Lc; 22.1106 s-1 in our operating conditions), and g,f is the gain or filter bandwidth, depending
on the intracavity elements [15]. The single-mode characteristic time tc is plotted in figure 8 as a function of the cavity
length, with (f = 1.1 nm FWHM – parabolic fit) and without (g = 34 nm) the 50-µm Fabry-Perot etalon inserted.
Figure 8: Characteristic generation time tc for an ideal homogeneous laser to collapse to
single-mode operation, with and without a 50µm thick etalon filter in the cavity; non-linear
mode coupling is neglected. Dot line #1: acoustic or/and thermal noise typical perturbation
time; dot line #2: multimode pump fluctuations perturbation time. The top arrow shows the
actual cavity length used in the experiment.
In our experimental conditions, without the etalon filter inside the laser cavity, the laser emission was not single-mode
and the cw (stationary) laser bandwidth L covered about several hundreds cavity modes. From (Eq. 4) it corresponds to
a generation time - the duration necessary for the laser spectrum to narrow into a single mode before a perturbation arises
- of about 100 µs. Since this value is one order of magnitude shorter than the typical characteristic time of acoustic
or/and thermal noise (in the millisecond range [3, 13]), we believe that the multimode nature of our fiber-coupled pump
laser diode generates strong pump intensity fluctuations at a time scale shorter than 100 µs, which in turns destabilize the
central wavelength of the gain spectrum and interrupt the spectral narrowing of the VECSEL. From the simulations in
figure 8 it is obvious that even our low-finesse etalon, with a FSR of 2 THz ( ~ 7 nm @ 1003 nm) which covers
numerous longitudinal modes, helps to reduce the characteristic time of the single-mode operation below the typical
fluctuations time of the set-up induced by acoustic, thermal noise and/or pump-induced gain fluctuations. It allows thus
to obtain a strong single-mode operation, as observed experimentally.
VI. INTRACAVITY SECOND HARMONIC GENERATION AT 501.7 NM
A four-mirror Z-cavity with a total length of around 630 mm was implemented for intracavity second harmonic
generation at 501.7 nm (figure 9). Mirrors M 1 and M2 had a radius of curvature of 200 mm, and a reflectivity > 99.98 %
between 900-1100 nm. Mirror M3 was a 75 mm spherical mirror with a reflectivity > 99.9 % in the range 650-1100 nm.
The laser cavity has two different outputs with similar powers at 501.7 nm because of the high transmission of all the
mirrors in the visible range. A type-I temperature-tuned non-critical phase-matching KNbO3 crystal was placed inside an
oven at a temperature of 75.6 °C. The nonlinear crystal was b-cut, 9.5 mm-long and anti-reflection coated for both
fundamental and second harmonic wavelengths. The waist radius inside the KNbO3 crystal was evaluated to about 90
8/11
Proceedings Photonics Europe Strasbourg 2006
µm. A Lyot filter and a 100 µm-thick FP etalon were also inserted not only to select the desire emission wavelength, but
also to obtain a stable single frequency operation at both wavelengths 1003.4 nm and 501.7 nm, which was impossible to
achieve without the etalon. The crystal was oriented such as the intracavity fundamental laser beam at 1003.4 nm was
polarized along the a-axis of the non-linear crystal, while the second harmonic generated beam was polarized along the
c-axis.
Figure 9: Experimental set-up achieved for the second harmonic generation process at 501.7 nm in a
four- mirror cavity with the Lyot Filter, the etalon and the oven for the KNbO 3 doubling crystal.
Figure 10 represents the single-frequency output power of the green-blue emission as function of the incident pump
power at 808 nm. A total maximum cw output power of 62 mW at 501.7 nm in two beams for a 6.5 W pump power was
obtained. The infrared intracavity power was estimated to be about 10 W, leading to a second harmonic generation
conversion efficiency of only 0.6 %. This low conversion efficiency can be attributed to the constraints imposed by the
oven dimensions, which lead to a rather large waist inside the crystal and a confocal parameter much longer (~ 44 mm)
than the crystal length [15]. Moreover, the residual losses introduced by the intracavity elements and the mirrors reduce
the available infrared laser power. The frequency spectrum of the fundamental infrared laser beam was also analysed
with the same Fabry-Perot interferometer as previously. The inset of figure 10 shows a typical scanning trace, which
demonstrates the single-frequency operation of the laser. Though these results are very preliminary and the second
harmonic generation process is not yet fully optimized as explained, the reached single frequency output power at 501.7
nm is similar to the one obtained with an Yb-doped laser crystal and a KNbO3 frequency-doubling crystal by Jacquemet
et al. [16], with the advantage of a simpler design of the laser cavity. Therefore, considering these as previous
measurements, future prospects focused on the enhancement of the SHG, on the stabilization of the laser emission and on
the change of the doubling crystal are going to be carried out.
9/11
Proceedings Photonics Europe Strasbourg 2006
Figure 10: CW laser output power at = 501.7 nm vs the incident pump for a 75.6 °C KNbO3 crystal
temperature. The inset shows the frequency spectrum confirming single-frequency behavior in the
blue-green.
VII. CONCLUSION
In conclusion, in this work the highest output power in a single-frequency operation of a diode-pumped VECSEL has
been reported. Thanks to an original technique based on the AuIn 2 solid liquid inter diffusion bonding, the active
structure has been bonded on a SiC substrate exhibiting a low thermal impedance. The AuIn 2 bonding allows further
processing steps at elevated temperatures and does not induce any strain-related degradation of the optical properties of
the structure. We then obtained a maximum output power of 1.7 W in a simple and compact plano-concave cavity,
limited by the available incident pump power. The improvement of the thermal dissipation from the active region has
been experimentally checked with both laser characterizations and temperature mappings under high power optical
pumping. It was theoretically explained through the evaluation of the thermal resistance of the ½ -VCSEL, which allows
us to predict the conditions for optimal laser operation.
With the insertion of only a solid Fabry-Perot etalon it has been possible to pick the emission wavelength at
1003 nm and to force a single-frequency operation with up to 500 mW of output power. A simple model has been used to
explain that laser emission on a single longitudinal mode is possible even with our low finesse and large FSR solid
etalon. We further demonstrate that, due to the multimode nature of the optical pump source used, an intracavity filter is
necessary to develop high power single-frequency VECSEL systems.
Finally, motivated by the interest of blue-green single-frequency emission around 501.7 nm for metrological
applications such as iodine-based laser stabilization, second harmonic generation had been performed with a KNbO3
nonlinear crystal. A maximum visible power of 62 mW has been generated in a four-mirror Z-cavity. These results can
be compared to those obtained with bulk ytterbium-doped crystals in a more complicated architecture.
REFERENCES
1.
M. Kuznetsov, F. Hakimi, R. Sprague and A. Mooradian, "Design and characteristics of high-power (>0.5-W CW)
diode-pumped vertical-external-cavity surface-emitting semiconductor lasers with circular TEM00 beams," IEEE J.
Selec. Topics in Quantum Electron. 5, 561 (1999).
10/11
Proceedings Photonics Europe Strasbourg 2006
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
W. J. Alford, T. D. Raymond and A. A. Allerman, "High power and good beam quality at 980 nm from a vertical
external-cavity surface-emitting laser," J. Opt. Soc. Am. B. 19, 663 (2002).
S. Lutgen, T. Albrecht, P. Brick, W. Reill, J. Luft and W. Späth, "8-W high-efficicency continuous-wave
semiconductor disk laser at 1000 nm," App Phys Lett 82, pp 3620-3622 (2003)
R. H. Abram, K. S. Gardner, E. Riis and A. I. Ferguson, “Narrow linewidth operation of a tunable optically pumped
semiconductor laser,” Opt. Exp. 12, 5434 (2004).
J. E. Hastie, J. M. Hopkins, S. Calvez, C. W. Jeon, D. Burns, R. H. Abram, E. Riis, A. I. Ferguson and M. D.
Dawson, “0.5-W single transverse-mode operation of an 850 nm diode-pumped surface-emitting semiconductor
laser,” IEEE Photon. Technol. Lett. 13, 894 (2003).
F. Du Burck, C. Daussy, A. Amy-Klein, A. Goncharov, O. Lopez, and C. Chardonnet, “Frequency Measurement of
an Ar+ Laser Stabilized on Narrow Lines of Molecular Iodine at 501.7 nm,” IEEE J.Trans. Intrum. and Meas., 54,
754-758 (2005)
W. Y. Cheng, L. Chen, T. H. Yoon, J. L. Hall and Y. Ye, "Sub-doppler molecular-iodine transitions near
dissociation limit (523-498 nm)," Opt. Lett. 27, 571 (2002).
J. Dion, I. Sagnes and M. Strassner, “High output power GaAs-based Vertical External Cavity Surface Emitting
Lasers achieved by AuIn2 Solid Liquid Inter Diffusion Bonding”, State-of-the-Art Program on Compound
Semiconductors XLI (2004)
L.A.Coldren, S.W. Morzine, Diode lasers and photonic integrated circuits, chapter IV, 1995, Wiley Interscience
L. Bernstein, ``Semiconductor joining by SLID Process, I. The systems Ag-In, Au-In and Cu-In,'' J. Electrochem.
Soc., 113, 1282-1288 (1966)
M.Reichling and H. Grönbeck, “Harmonic heat flow in isotropic layered systems and its use for thin film thermal
conductivity measurements”, J.Appl.Phys. 75, 1914-1922 (1994).
S. Chénais, S. Forget, F. Druon, F. Balembois and P. Georges, “Direct and absolute temperature mapping and heat
transfer measurements in diode-end-pumped Yb:YAG”, Appl.Phys.,B 79 p.221 ( 2004)
A. Ouvrard, A. Garnache, L. Cerutti, F. Genty, D. Romanini, “Single Frequency Tunable Sb–based VCSELs
emitting at 2.3 µm”, IEEE Photon. Techn. Lett.,17, 2020-2022 (2005)
A. Garnache, A. Ouvrard and D. Romanini, “Spectro-temporal dynamics of external-cavity VCSELs: SingleFrequency operation” Proc. IEEE CLEO Europe 2005, paper CB19
A. Garnache, A. Kachanov, F. Stoeckel and R. Houdre, “Diode-pumped broadband vertical external cavity surface
emitting semiconductor laser applied to high-sensitivity intracavity absorption spectroscopy,” J. Opt. Soc. Am. B 7,
1589 (1999).
M. Jacquemet, F. Druon, F. Balembois, P. Georges, and B. Ferrand, “Blue-green single-frequency laser based on
intracavity frequency doubling of a diode-pumped Ytterbium-doped laser,” Optics Express 13, 2345-2350, (2005).
11/11